Data Imputation Techniques Research Articles

Machine Learning in Cardiac Surgery: Predicting Mortality and Readmission.

Predicting outcomes in open-heart surgery can be challenging. Unexpected readmissions, long hospital stays, and mortality have economic implications. In this study, we investigated machine learning (ML) performance in data visualization and predicting patient outcomes associated with open-heart surgery. We evaluated 8,947 patients who underwent cardiac surgery from April 2006 to January 2018. Data visualization and classification were performed at cohort-level and patient-level using clustering, correlation matrix, and seven different predictive models for predicting three outcomes ("Discharged," "Died," and "Readmitted") at binary level. Cross-validation was used to train and test each dataset with the application of hyperparameter optimization and data imputation techniques. Machine learning showed promising performance for predicting mortality (AUC 0.83 ± 0.03) and readmission (AUC 0.75 ± 0.035). The cohort-level analysis revealed that ML performance is comparable to the Society of Thoracic Surgeons (STS) risk model even with limited number of samples ( e.g. , less than 3,000 samples for ML versus more than 100,000 samples for the STS risk models). With all cases (8,947 samples, referred as patient-level analysis), ML showed comparable performance to what has been reported for the STS models. However, we acknowledge that it remains unknown at this stage as to how the model might perform outside the institution and does not in any way constitute a comparison of the performance of the internal model with the STS model. Our study demonstrates a systematic application of ML in analyzing and predicting outcomes after open-heart surgery. The predictive utility of ML in cardiac surgery and clinical implications of the results are highlighted.

Missing Data Imputation in the Internet of Things Sensor Networks

The Internet of Things (IoT) has had a tremendous impact on the evolution and adoption of information and communication technology. In the modern world, data are generated by individuals and collected automatically by physical objects that are fitted with electronics, sensors, and network connectivity. IoT sensor networks have become integral aspects of environmental monitoring systems. However, data collected from IoT sensor devices are usually incomplete due to various reasons such as sensor failures, drifts, network faults and various other operational issues. The presence of incomplete or missing values can substantially affect the calibration of on-field environmental sensors. The aim of this study is to identify efficient missing data imputation techniques that will ensure accurate calibration of sensors. To achieve this, we propose an efficient and robust imputation technique based on k-means clustering that is capable of selecting the best imputation technique for missing data imputation. We then evaluate the accuracy of our proposed technique against other techniques and test their effect on various calibration processes for data collected from on-field low-cost environmental sensors in urban air pollution monitoring stations. To test the efficiency of the imputation techniques, we simulated missing data rates at 10–40% and also considered missing values occurring over consecutive periods of time (1 day, 1 week and 1 month). Overall, our proposed BFMVI model recorded the best imputation accuracy (0.011758 RMSE for 10% missing data and 0.169418 RMSE at 40% missing data) compared to the other techniques (kNearest-Neighbour (kNN), Regression Imputation (RI), Expectation Maximization (EM) and MissForest techniques) when evaluated using different performance indicators. Moreover, the results show a trade-off between imputation accuracy and computational complexity with benchmark techniques showing a low computational complexity at the expense of accuracy when compared with our proposed technique.

Data imputation in deep neural network to enhance breast cancer detection

AbstractBreast cancer is one of the most precarious cancers that claims many women's' lives every year. The existing automated systems for mammography datasets are designed to detect the abnormalities and classify them as benign or malignant. However, these systems ignore the incompleteness of data due to missing or irrelevant values in the dataset, which severely impede the automated system's reliability. To enhance the reliability of the decision outcomes, this study proposes a composite imputed multilayer fully connected deep networks (MFCN) that can replace the incomplete values in the dataset to augment the efficient feature learning for breast cancer classification. The data imputation techniques and one‐hot encoding transformation applied to the dataset are proved to improve the feature representation. The imputed features extracted from the instances are fed into the MFCN to classify malignant and benign by utilizing mammographic mass (MM) and INbreast datasets. The performance of the proposed missing data imputation treated MFCN (MdI‐MFCN) is compared with competitive machine learning (ML) classifiers by incorporating stratified 6‐fold cross‐validation. The experimental results of the models are evaluated with and without imputed data using the five performance metrics, namely, accuracy, precision, recall, F1‐score, and specificity. The performance of the proposed MdI‐MFCN has outperformed the competitive classifiers with the highest accuracy of 93% and 98% of the area under the receiver operating characteristic (ROC) curve on MM and INbreast datasets, respectively. The simplicity and reliability of this proposed approach its potential in characterizing mammography datasets during breast cancer diagnosis.

Systematic Review on Missing Data Imputation Techniques with Machine Learning Algorithms for Healthcare

Missing data is one of the most common issues encountered in data cleaning process especially when dealing with medical dataset. A real collected dataset is prone to be incomplete, inconsistent, noisy and redundant due to potential reasons such as human errors, instrumental failures, and adverse death. Therefore, to accurately deal with incomplete data, a sophisticated algorithm is proposed to impute those missing values. Many machine learning algorithms have been applied to impute missing data with plausible values. However, among all machine learning imputation algorithms, KNN algorithm has been widely adopted as an imputation for missing data due to its robustness and simplicity and it is also a promising method to outperform other machine learning methods. This paper provides a comprehensive review of different imputation techniques used to replace the missing data. The goal of the review paper is to bring specific attention to potential improvements to existing methods and provide readers with a better grasps of imputation technique trends.

Robust estimation of traffic density with missing data using an adaptive-R extended Kalman filter

Traffic density is a crucial indicator of traffic congestion, but measuring it directly is often infeasible and hence, it is usually estimated based on other measurements. However, a challenge in measuring traffic parameters is the high probability of sensor failure, which results in missing measurement or missing data. To overcome this difficulty, in this paper, we propose a novel adaptive-R extended Kalman filter (AREKF) combined with a model-based data imputation technique to estimate traffic density. We show analytically that the AREKF is able to accurately estimate the density even when the noise covariance matrices are not accurately known. Microscopic traffic simulations demonstrated the efficacy of the AREKF, where the estimated density is fed into a real-time ramp metering control algorithm to control vehicle flow on a merging road, which is highly susceptible to traffic congestion. The results show that the proposed AREKF with data imputation is able to accurately estimate the traffic density even when data is missing, and the ramp-metering controller significantly improves the traffic flow and thus, alleviates congestion.

Application of machine learning missing data imputation techniques in clinical decision making: taking the discharge assessment of patients with spontaneous supratentorial intracerebral hemorrhage as an example

BackgroundThere are often many missing values in medical data, which directly affect the accuracy of clinical decision making. Discharge assessment is an important part of clinical decision making. Taking the discharge assessment of patients with spontaneous supratentorial intracerebral hemorrhage as an example, this study adopted the missing data processing evaluation criteria more suitable for clinical decision making, aiming at systematically exploring the performance and applicability of single machine learning algorithms and ensemble learning (EL) under different data missing scenarios, as well as whether they had more advantages than traditional methods, so as to provide basis and reference for the selection of suitable missing data processing method in practical clinical decision making.MethodsThe whole process consisted of four main steps: (1) Based on the original complete data set, missing data was generated by simulation under different missing scenarios (missing mechanisms, missing proportions and ratios of missing proportions of each group). (2) Machine learning and traditional methods (eight methods in total) were applied to impute missing values. (3) The performances of imputation techniques were evaluated and compared by estimating the sensitivity, AUC and Kappa values of prediction models. (4) Statistical tests were used to evaluate whether the observed performance differences were statistically significant.ResultsThe performances of missing data processing methods were different to a certain extent in different missing scenarios. On the whole, machine learning had better imputation performance than traditional methods, especially in scenarios with high missing proportions. Compared with single machine learning algorithms, the performance of EL was more prominent, followed by neural networks. Meanwhile, EL was most suitable for missing imputation under MAR (the ratio of missing proportion 2:1) mechanism, and its average sensitivity, AUC and Kappa values reached 0.908, 0.924 and 0.596 respectively.ConclusionsIn clinical decision making, the characteristics of missing data should be actively explored before formulating missing data processing strategies. The outstanding imputation performance of machine learning methods, especially EL, shed light on the development of missing data processing technology, and provided methodological support for clinical decision making in presence of incomplete data.

Missing data imputation on biomedical data using deeply learned clustering and L2 regularized regression based on symmetric uncertainty

Big data era in healthcare led to the generation of high dimensional datasets like genomic datasets, electronic health records etc. One among the critical issues to be addressed in such datasets is handling incomplete data that may yield misleading results if not handled properly. Imputation is considered to be an effective way when the missing data rate is high. While imputation accuracy and classification accuracy are the two important metrics generally considered by most of the imputation techniques, high dimensional datasets such as genomic datasets motivated the need for imputation techniques that are also computationally efficient and preserves the structure of the dataset. This paper proposes a novel approach to missing data imputation in biomedical datasets using an ensemble of deeply learned clustering and L2 regularized regression based on symmetric uncertainty. The experiments are conducted with different proportion of missing data on both genomic and non-genomic biomedical datasets for different types of missingness pattern. Our proposed approach is compared with seven proven baseline imputation methods and two recently proposed imputation approaches. The results show that the proposed approach outperforms the other approaches considered in our experimentation in terms of imputation accuracy and computational efficiency despite preserving the structure of the dataset. Thus, the overall classification accuracy of the biomedical classification tasks is also improved when our proposed missing data imputation technique is used.

Performance Analysis of Conventional Machine Learning Algorithms for Identification of Chronic Kidney Disease in Type 1 Diabetes Mellitus Patients.

Chronic kidney disease (CKD) is one of the severe side effects of type 1 diabetes mellitus (T1DM). However, the detection and diagnosis of CKD are often delayed because of its asymptomatic nature. In addition, patients often tend to bypass the traditional urine protein (urinary albumin)-based CKD detection test. Even though disease detection using machine learning (ML) is a well-established field of study, it is rarely used to diagnose CKD in T1DM patients. This research aimed to employ and evaluate several ML algorithms to develop models to quickly predict CKD in patients with T1DM using easily available routine checkup data. This study analyzed 16 years of data of 1375 T1DM patients, obtained from the Epidemiology of Diabetes Interventions and Complications (EDIC) clinical trials directed by the National Institute of Diabetes, Digestive, and Kidney Diseases, USA. Three data imputation techniques (RF, KNN, and MICE) and the SMOTETomek resampling technique were used to preprocess the primary dataset. Ten ML algorithms including logistic regression (LR), k-nearest neighbor (KNN), Gaussian naïve Bayes (GNB), support vector machine (SVM), stochastic gradient descent (SGD), decision tree (DT), gradient boosting (GB), random forest (RF), extreme gradient boosting (XGB), and light gradient-boosted machine (LightGBM) were applied to developed prediction models. Each model included 19 demographic, medical history, behavioral, and biochemical features, and every feature’s effect was ranked using three feature ranking techniques (XGB, RF, and Extra Tree). Lastly, each model’s ROC, sensitivity (recall), specificity, accuracy, precision, and F-1 score were estimated to find the best-performing model. The RF classifier model exhibited the best performance with 0.96 (±0.01) accuracy, 0.98 (±0.01) sensitivity, and 0.93 (±0.02) specificity. LightGBM performed second best and was quite close to RF with 0.95 (±0.06) accuracy. In addition to these two models, KNN, SVM, DT, GB, and XGB models also achieved more than 90% accuracy.

EFFECTS OF DIFFERENT MULTIPLE IMPTUTATION TECHNIQUES ON THE MODEL FIT OF CONFIRMATORY FACTOR ANALYSIS

So far, many researches have been conducted to investigate the impact of missing data on statistical analysis and various methods have been developed to deal with the problem. The methods based on removing observations with missing values from the dataset cause the sample size to drop dramatically and the statistical power of the analyzes to be decreased. Therefore, as an alternative solution, the estimation of missing values seized intensive attention of researchers. Among these methods, multiple imputation techniques are relatively more recent and provide better estimations. Considering the superiority of multiple imputation techniques, the aim of the current study is to investigate the effects of different multiple imptutation techniques on the model fit of confirmatory factor analysis. For this aim, datasets with the unidimensional structure were simulated to manipulate sample size, missing data mechanism, percentage of missing data, number of items and missing data imputation technique. The effect of multiple imputation techniqes was evaluated based on the difference of 𝜒² model fit statistics for complete datasets and imputed datasets. The results showed that, multiple impuation techniques provided better results than conventional regression based imputation. Those finding were discussed later and some recommendations were given for better testing applications.

LR Hunting: A Random Forest Based Cell-Cell Interaction Discovery Method for Single-Cell Gene Expression Data.

Cell–cell interactions (CCIs) and cell–cell communication (CCC) are critical for maintaining complex biological systems. The availability of single-cell RNA sequencing (scRNA-seq) data opens new avenues for deciphering CCIs and CCCs through identifying ligand-receptor (LR) gene interactions between cells. However, most methods were developed to examine the LR interactions of individual pairs of genes. Here, we propose a novel approach named LR hunting which first uses random forests (RFs)-based data imputation technique to link the data between different cell types. To guarantee the robustness of the data imputation procedure, we repeat the computation procedures multiple times to generate aggregated imputed minimal depth index (IMDI). Next, we identify significant LR interactions among all combinations of LR pairs simultaneously using unsupervised RFs. We demonstrated LR hunting can recover biological meaningful CCIs using a mouse cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq) dataset and a triple-negative breast cancer scRNA-seq dataset.

Hybrid diabetes disease prediction framework based on data imputation and outlier detection techniques

AbstractIn the field of medical science, accurate prediction is a difficult and challenging task. But, the presence of missing values and outliers can make the prediction task more complicated. Many researchers address the issue of missing value in medical data, either detect the missing value and delete the respective data instances from the dataset or adopt some default methods such as mean, median, neighbour etc., for filling the missing value. However, both methods are lacking to produce optimal results. Furthermore, outliers are also presented in data and degraded the performance of classifier. Few researchers also focus on the outlier detection in medical dataset, but it is not fully explored till date. This work considers the two well‐known problems of data that is, (i) missing value imputation, and (ii) outlier. The missing value imputation issue is addressed through K‐Mean++ based data imputation technique. This technique also validates the data through clustering and also compute the values for missing data. The outlier can be detected through an ABC based outlier detection technique. Further, the final outcome is determined using LS‐SVM classifiers. Hence, this work presents a hybrid disease diagnosis framework for diabetes prediction, called hybrid diabetes prediction framework. The reason behind to choose the diabetes dataset for implementation as it contains 763 missing values and several outliers. The simulation results showed that proposed hybrid framework effectively determines the missing values and outliers in diabetes dataset. Further, the performance of proposed hybrid diabetes prediction framework is evaluated using accuracy, sensitivity, specificity, kappa and AUC parameters and compared with 34 state of art techniques. Results confirmed that proposed hybrid framework obtains 96.57%, 93.37%, 98.12%, 98.17%, and 95.43% accuracy, sensitivity, specificity, kappa and AUC rate respectively.

Forecasts of tropospheric ozone in the Metropolitan Area of Rio de Janeiro based on missing data imputation and multivariate calibration techniques.

Multivariate calibration based on partial least squares, random forest, and support vector machine methods, combined with the MissForest imputation algorithm, was used to understand the interaction between ozone and nitrogen oxides, carbon monoxide, wind speed, solar radiation, temperature, relative humidity, and others, the data of which were collected by air quality monitoring stations in the metropolitan area of Rio de Janeiro in four distinct sites between, 2014 and, 2018. These techniques provide an easy and feasible way of modeling and analyzing air pollutants and can be used when coupled with other methods. The results showed that random forest and support vector machine chemometric techniques can be used in modeling and predicting tropospheric ozone concentrations, with a coefficient of determination for making predictions up to 0.92, a root-mean square error of calibration between 4.66 and 27.15µgm-3, and a root-mean square error of prediction between 4.17 and 22.45µgm-3, depending on the air quality monitoring stations and season.

Comparison of machine learning techniques for the identification of human activities from inertial sensors available in a mobile device after the application of data imputation techniques

Human activity recognition (HAR) is a significant research area due to its wide range of applications in intelligent health systems, security, and entertainment games. Over the past few years, many studies have recognized human daily living activities using different machine learning approaches. However, the performance of a machine learning algorithm varies based on the sensing device type, the number of sensors in that device, and the position of the underlying sensing device. Moreover, the incomplete activities (i.e., data captures) in a dataset also play a crucial role in the performance of machine learning algorithms. Therefore, we perform a comparative analysis of eight commonly used machine learning algorithms in different sensor combinations in this work. We used a publicly available mobile sensors dataset and applied the k-Nearest Neighbors (KNN) data imputation technique for extrapolating the missing samples. Afterward, we performed a couple of experiments to figure out which algorithm performs best at which sensors' data combination. The experimental analysis reveals that the AdaBoost algorithm outperformed all machine learning algorithms for recognizing five different human daily living activities with both single and multi-sensor combinations. Furthermore, the experimental results show that AdaBoost is capable to correctly identify all the activities presented in the dataset with 100% classification accuracy.

A novel framework for imputing large gaps of missing values from time series sensor data of marine machinery systems

ABSTRACT Condition-based maintenance is a maintenance strategy that implements Industrial Internet of Things to monitor the assets’ condition. Despite its undeniable benefits, several challenges are encountered, such as the incompleteness of sensor data. Hence, while data imputation is an important practise, there is a lack of analysis and formalisation of data imputation in the maritime industry. Accordingly, a novel framework is introduced by implementing the first-order Markov chain in tandem with a multivariate imputation approach based on a comparative methodology of 16 machine learning and time series forecasting models. To highlight its performance efficiency, a comparative study is conducted between the proposed framework and the MICE approach by the implementation of a case study on a total of 4 parameters, obtained from sensors installed on the marine machinery systems of a cargo vessel. The results demonstrated an improvement of 21–77%, indicating its performance efficiency as a data imputation technique.

Kernel weighted least square approach for imputing missing values of metabolomics data

Mass spectrometry is a modern and sophisticated high-throughput analytical technique that enables large-scale metabolomic analyses. It yields a high-dimensional large-scale matrix (samples × metabolites) of quantified data that often contain missing cells in the data matrix as well as outliers that originate for several reasons, including technical and biological sources. Although several missing data imputation techniques are described in the literature, all conventional existing techniques only solve the missing value problems. They do not relieve the problems of outliers. Therefore, outliers in the dataset decrease the accuracy of the imputation. We developed a new kernel weight function-based proposed missing data imputation technique that resolves the problems of missing values and outliers. We evaluated the performance of the proposed method and other conventional and recently developed missing imputation techniques using both artificially generated data and experimentally measured data analysis in both the absence and presence of different rates of outliers. Performances based on both artificial data and real metabolomics data indicate the superiority of our proposed kernel weight-based missing data imputation technique to the existing alternatives. For user convenience, an R package of the proposed kernel weight-based missing value imputation technique was developed, which is available at https://github.com/NishithPaul/tWLSA.

Robustness to Missing Features using Hierarchical Clustering with Split Neural Networks (Student Abstract)

The problem of missing data has been persistent for a long time and poses a major obstacle in machine learning and statistical data analysis. Past works in this field have tried using various data imputation techniques to fill in the missing data, or training neural networks (NNs) with the missing data. In this work, we propose a simple yet effective approach that clusters similar input features together using hierarchical clustering and then trains proportionately split neural networks with a joint loss. We evaluate this approach on a series of benchmark datasets and show promising improvements even with simple imputation techniques. We attribute this to learning through clusters of similar features in our model architecture.

Learning from Imbalanced Educational Data Using Ensemble Machine Learning Algorithms

Student’s academic performance is one of the most important parameters for evaluating the standard of any institute. It has become a paramount importance for any institute to identify the student at risk of underperforming or failing or even drop out from the course. Machine Learning techniques may be used to develop a model for predicting student’s performance as early as at the time of admission. The task however is challenging as the educational data required to explore for modelling are usually imbalanced. We explore ensemble machine learning techniques namely bagging algorithm like random forest (rf) and boosting algorithms like adaptive boosting (adaboost), stochastic gradient boosting (gbm), extreme gradient boosting (xgbTree) in an attempt to develop a model for predicting the student’s performance of a private university at Meghalaya using three categories of data namely demographic, prior academic record, personality. The collected data are found to be highly imbalanced and also consists of missing values. We employ k-nearest neighbor (knn) data imputation technique to tackle the missing values. The models are developed on the imputed data with 10 fold cross validation technique and are evaluated using precision, specificity, recall, kappa metrics. As the data are imbalanced, we avoid using accuracy as the metrics of evaluating the model and instead use balanced accuracy and F-score. We compare the ensemble technique with single classifier C4.5. The best result is provided by random forest and adaboost with F-score of 66.67%, balanced accuracy of 75%, and accuracy of 96.94%.

Investigating the impact of missing data imputation techniques on battery energy management system

Effective control of energy storage system (ESS), supplying an ancillary service to a grid, requires effective and critical calculation of state-of-charge (SoC). Charging and discharging values from battery operations are essential in calculating the efficiency and performance of a storage system. This information can also be a key to understand and forecast peak demand performance. Missing data is a real problem in any operations system, and it appears to be more common within powers systems due to sensor and/or network malfunctioning problems. Missing data imputation techniques have evolved in power systems research using smart meter data, but little research has gone into understanding how missing data can be best handled within storage management systems. This paper builds on a year's worth of charging and discharging data collected from a real 6MW/10MWh lithium-ion storage battery deployed on the distribution network at Leighton Buzzard, UK. Using R Studio version (1.3.959-1) open-source software, eight selected imputation techniques were applied in identifying the best suited technique in replacing various missing data amounts and patterns. Findings from the study open up avenues for discussion and debate in identifying an appropriate imputation technique within the storage management context. The study also provides a pioneering lead in understanding the importance of decomposition in evaluating the right imputation technique.

Fuzzy based Techniques for Handling Missing Values

Usually, time series data suffers from high percentage of missing values which is related to its nature and its collection process. This paper proposes a data imputation technique for imputing the missing values in time series data. The Fuzzy Gaussian membership function and the Fuzzy Triangular membership function are proposed in a data imputation algorithm in order to identify the best imputation for the missing values where the membership functions were used to calculate weights for the data values of the nearest neighbor’s before using them during imputation process. The evaluation results show that the proposed technique outperforms traditional data imputation techniques where the triangular fuzzy membership function has shown higher accuracy than the gaussian membership function during evaluation.

Fuzzy C-mean Missing Data Imputation for Analogy-based Effort Estimation

The accuracy of effort estimation in one of the major factors in the success or failure of software projects. Analogy-Based Estimation (ABE) is a widely accepted estimation model since its flow human nature in selecting analogies similar in nature to the target project. The accuracy of prediction in ABE model in strongly associated with the quality of the dataset since it depends on previous completed projects for estimation. Missing Data (MD) is one of major challenges in software engineering datasets. Several missing data imputation techniques have been investigated by researchers in ABE model. Identification of the most similar donor values from the completed software projects dataset for imputation is a challenging issue in existing missing data techniques adopted for ABE model. In this study, Fuzzy C-Mean Imputation (FCMI), Mean Imputation (MI) and K-Nearest Neighbor Imputation (KNNI) are investigated to impute missing values in Desharnais dataset under different missing data percentages (Desh-Miss1, Desh-Miss2) for ABE model. FCMI-ABE technique is proposed in this study. Evaluation comparison among MI, KNNI, and (ABE-FCMI) is conducted for ABE model to identify the suitable MD imputation method. The results suggest that the use of (ABE-FCMI), rather than MI and KNNI, imputes more reliable values to incomplete software projects in the missing datasets. It was also found that the proposed imputation method significantly improves software development effort prediction of ABE model.